Re: [Miro-devel] Re: Re: [Playexrstage-developers] Common robotics data exchange format...?

Herman/Brian/Carle:
  For my meager 2 cents, while the development of standards won't be
easy, they are a worthy goal to pursue (look at the success of the
internet/websites/browsers, all of which is possible due to agreed upon
standards and of course one could easily argue that it was a much
simpler domain to work in and hence easy to achieve success). 


On Mon, 2004-08-16 at 09:26, Herman Bruyninckx wrote:
> On Mon, 16 Aug 2004, Carle [iso-8859-1] Côté wrote:
> 
> [...]
> > We have created an XML data representation of this library as we're
> > now using this format for data streaming
> > (http://marie.sourceforge.net/en/Documentation.html#DATATYPES). It
is a very
> > preliminary work, but it might be useful for discussion. Note that
we're
> > supposed to move to the same data representation as Player for
datatypes that
> > can be shared, as it is "de facto" standard.
> 
> All this is good news :-)
> 
> But most of the data types mentioned in the cited on-line document
are,
> in my opinion, beyond the scope of what I had in mind: they do not
> represent states of the robots, but of (less standardizable)
peripherals
> and control modes.

I agree with Herman here.  All robots have states and they are the
obvious place to start.  Another reason for concentrating on states
would be its use throughout the system. The states will (depending on
the robot) be used in numerous places and thus woven into the "fabric"
of the robot. This dependency makes them difficult "change" once they
have initially been used and hence their is a large benefit for
standardization.
I would also add all robots have a mean of changing their states, ie
commands for movement, velocity etc, and that "standards" in this area
would also be beneficial (I don't think this is what Herman refers to as
a "control mode").

For discussion purposes where do others see the greatest benefits being
accrued from the use of standards?  Herman makes an argument for the
robot state.  I added the change in robot state.   

Would it be reasonable to argue that a greater benefit is accrued from
the standardization of "abstracts" such as states over the
standardization of interfaces to peripheral/devices?
 
 

> 
> Concretely speaking, I would start from the file
>   client_libs/c++/playerclient.h
> in the player source tree, and discuss its basic "proxy" data
structures:
> descartesproxy, positionproxy, and position3dproxy. I would then add
> "manipulator" and "humanoid" stuff. And that's it. (To start with.)
> 
> Herman

Greg B.




-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285